Toitumuse visuaalse ja automaatse hindamise võrdlus
Laen...
Kuupäev
2022
Kättesaadav alates
10.09.2022
Autorid
Matto, Anette
Ajakirja pealkiri
Ajakirja ISSN
Köite pealkiri
Kirjastaja
Eesti Maaülikool
Abstrakt
Bakalaureusetöö on osa söötmisteaduste õppetooli täidetavast projektist „Toitumuse
automaatne hindamine ja biosensorid piimakarja haldamisel“. Teadusuuringuid, kus
võrreldakse visuaalset ja automaatset toitumuse hindamise süsteemi, on maailmas vähe tehtud
ning Eestis pole teadaolevalt varem sarnast uurimust läbi viidud. Töö peamisteks
eesmärkideks oli võrrelda kahe hindaja poolt pandud toitumushinde kokkulangevust ja
võrrelda toitumuse automaatse hindamise süsteemi ning visuaalse toitumushinde
kokkulangevust. Valimisse võeti 468 Kehtna Mõisa OÜ eesti holsteini tõugu lehma
toitumushinnet. Toitumust hinnati viiel korral ajavahemikus 2021. a detsember kuni 2022. a
jaanuar 5-pallisel skaalal 0,25-pallise sammuga. Visuaalseid hindeid võrreldi DeLavali
toitumuse automaatse hindamise süsteemi poolt registreeritud toitumushinnetega. Leiti, et
kaks hindajat hindavad lehmade toitumust sarnaselt, sest 98,1% juhtudest langesid
toitumushinded kokku või ei erinenud rohkem kui 0,25 palli ja toitumushinnete keskmine
erinevus oli 0. Seega üks hindaja ei hinnanud toitumust süsteemselt teisest hindajast erinevalt.
See on kooskõlas varasemate uurimistulemustega. Leiti, et toitumuse automaatse hindamise
süsteem hindab toitumuse üldist dünaamikat laktatsiooni jooksul visuaalse hindamisega
sarnaselt. Samas automaatse hindamise süsteem hindas võrreldes inimestega loomade
toitumuse keskmiselt suuremaks ehk loomi rasvunumaks, mida kõhnem oli lehm, seda
suurem oli erinevus. Saadud tulemused on uudsed, kuna üheski teises toitumuse automaatse hindamise süsteemi ja visuaalse hindamissüsteemi võrdluses pole süsteemseid visuaalse ja
automaatse hindamise erinevusi välja toodud.
This Bachelor’s thesis is a part of the project „Automated body condition scoring and biosensors in dairy herd management“ carried out by the Chair of Animal Nutrition. Only few publications are available on the evaluation of an automatic body condition scoring system and this is the first such study in Estonia. The aim of this Bachelor’s thesis was to evaluate the repeatability of visual body condition scoring and to compare the automatic body condition assessment system with visual scoring. In total 468 body condition scores of Estonian Holstein cows were assessed from December 2021 to January 2022. Body condition score was assessed on a 5-point scale with 0.25-point increments. The DeLaval’s automatic body condition assessment system provided the automatic body condition scores. The results revealed that the results of two observers were in agreement as in 98.1% of the cases the scores were identical or differed by not more than 0.25 and the average difference between observers’ scores was 0. Thus, one observer did not evaluate body condition score systematically differently from the other. This is in agreement with previously published results. It was found that the automatic system evaluates the overall dynamics similarly to visual scoring. However, the automatic system overestimated body condition scores compared to visual assessment, but the size of the error depended on the body condition score – the thinner the cow, the bigger the difference. The results of this Bachelor’s thesis are novel as such systematic differences between visual and automatic scoring have not been reported before.
This Bachelor’s thesis is a part of the project „Automated body condition scoring and biosensors in dairy herd management“ carried out by the Chair of Animal Nutrition. Only few publications are available on the evaluation of an automatic body condition scoring system and this is the first such study in Estonia. The aim of this Bachelor’s thesis was to evaluate the repeatability of visual body condition scoring and to compare the automatic body condition assessment system with visual scoring. In total 468 body condition scores of Estonian Holstein cows were assessed from December 2021 to January 2022. Body condition score was assessed on a 5-point scale with 0.25-point increments. The DeLaval’s automatic body condition assessment system provided the automatic body condition scores. The results revealed that the results of two observers were in agreement as in 98.1% of the cases the scores were identical or differed by not more than 0.25 and the average difference between observers’ scores was 0. Thus, one observer did not evaluate body condition score systematically differently from the other. This is in agreement with previously published results. It was found that the automatic system evaluates the overall dynamics similarly to visual scoring. However, the automatic system overestimated body condition scores compared to visual assessment, but the size of the error depended on the body condition score – the thinner the cow, the bigger the difference. The results of this Bachelor’s thesis are novel as such systematic differences between visual and automatic scoring have not been reported before.
Kirjeldus
Bakalaureusetöö
Loomakasvatuse õppekaval
Märksõnad
bakalaureusetööd, piimaveised, toitumushinne, visuaalne ja automaatne toitumuse hindamine